Sam Altman:
A lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we’re going to be able to run them on our computers, and so we’re not going to need all the stuff in the data center.
But from a global allocation perspective, actually it’s quite terrible to have everybody, like, have a bunch of the world’s limited supply of DRAM sitting in their own computer and being used five percent of the time.
Like you want maximum utilisation in the cloud if you’re trying to deal with this very supply-constrained environment,
And somehow many people’s intuition is the opposite of that
TL;DR Guy selling product: “Here’s why owning your own product instead of renting it from me is a terrible idea”
Haha yes! And it would so much more efficient if we just like, put all the workers in a big warehouse in between shifts. We can drug them so they can sleep in stacked boxes, and then drug them again so they are ready to work after exactly 8 hours. They will get so much more work done if they don’t have to ever leave the factory, especially if we give everyone ostomy bags and catheters so they can empty their waste once before bed!
It’s going to be so fucking efficient guys. Like, you have no fucking idea how much more efficient we could make society if we just removed the concept of individual agency and autonomy from our ethical repertoire!
Awesome! Let’s name this system Seriously Lean And Very Efficient Resource Ytilization!